Thursday, Aug. 28, 2014

In-Play Trading On Betfair

Betting Exchange TutorialAlthough, as I have indicated in other articles on this Web site, I am no great trader; preferring to stick to what I know (backing horses), there is one aspect of trading that does intrigue me, and which I believe has strong possibilities for success, and that is trading on the in-play market.

I sit and study the in-play market every day, and I am intrigued by the volatility that exists; clearly people greening up positions. In the early days when I just used to watch the computer screen, I thought the market was a reflection of the race. I now understand this is not completely the case, and in this fact some possibilities have to exist.

As with any trade we are looking to back and lay at a high and low price or lay and back at a low and high price. Here is the first difficulty. Should we lay first and hope the horse runs badly during some part of the race that will allow us to back at a higher price to a stake that will provide a profit whatever the outcome? Alternatively, should we back first in the hope that the horse takes a prominent position at some point during the race and we can place a lay bet to give us a profit whatever the outcome?

This is not as simple a decision as might first appear. If we lay first, and the horse shoots off in front, it may be impossible to get the saving back bet on. Alternatively, if we initially place a back bet and the horse falls at the first fence, we are pretty well stuffed.

In my experience, there is far more likelihood that a horse will, at some point in the race, be available at longer odds than those obtainable at the start of the race. Thus it would appear that the lay first option is the one most likely to succeed. It also has the added advantage that should your horse fall at the first fence, you can sit back and count your winnings. Certainly, there are going to be times when a selection makes every yard of the running and the price just gets shorter and shorter, but these appear to be few and far between.

Another aspect to take into consideration is the price of the horse we are going to trade in-play. I think this strategy is better suited to longer priced horses (without being ridiculous and risking too much), say in the 5.0 to 10.0 range. The rationale behind this is twofold. Firstly, as we are taking a lay position initially, we want selections that the general public feel are not going to win. Therefore, any slight mistakes by the horse, any evidence of them being ridden along, or coming under pressure, will be a reinforcement of the initial opinion and will be reflected quickly in the market.

Secondly, the way Betfair is structured means that smaller movements in longer-priced horses will provide a greater differential than in shorter-priced selections. We need this differential in order to provide acceptable returns commensurate with the risk being taken.

Lets say we take a position that we are going to lay a horse at a price and back it in running once it moves two ticks from that price. For example, we lay a horse @ 5.3 to £100, in running we place a back bet for £96.36 @ 5.5. If the horse loses we win £3.64 and if it wins we take out £3.62. Those proponents of how easy it is to make money on the betting exchanges would no doubt say that this is easy money, but as always, we have to examine the reality.

If we lay a horse @ 5.3 to £100 with the intention of taking a back position @5.5 we have to consider what happens if the 5.5 is never attainable. The selection shoots off in front, jumps like a stag, the first and second favourite fall at the first fence. We are stuck with a liability of £430. Therefore, we have to accept that we are effectively risking £430 to win £3.64 massive odds on.

This bleak position can be tempered slightly with the frequency that the above scenario might happen; that is, the saving back bet doesnt get matched. There is a direct correlation between the chance of the back bet not getting matched (and us having to face the full liability) and the number of ticks (or the spread we are prepared to accept).

Suppose, for example we decide we are going to lay a horse @5.3 to £100, but back it @10 to £53. In this situation, our potential return is £47 regardless of whether the horse wins or loses. We have reduced the true odds to less than 10/1 on, but we have increased the likelihood of the bet not getting matched should the horse be prominent throughout (not just taking a long lead), and go on to win.

There is a stronger probability of a horse that subsequently wins moving from 5.3 to 5.5 in the market during running than there is of a horse that subsequently wins moving from 5.3 to 10. As with all things Betfair, you cant buck the odds.

It is clear that there should exist some magical nexus where the probability of getting matched is balanced against the probability of not getting matched. This would form the ultimate spread position.

Now, at this point, no doubt the old hands at this game are screaming about taking a stop loss. For those novices among you this is a method of minimising losses should the market turn against you. We will take the previous scenario of a £100 lay bet @ 5.3 with the intention of a £96.36 back bet @5.5.

We want to minimise our losses should things go against us, so we decide to set a stop loss position of 4.7 (that is should the horses odds touch 4.7 we will back it and take the loss). On paper this appears a reasonable option. If instead of placing the back bet @ 5.5 we place it at 4.7, then if the horse goes on to win we only lose £73.47 (if it loses we still win £3.64) as opposed to a potential loss of £430 if we dont get any sort of back bet on.

Firstly, we need to recognise that in order to make this work we need to employ some additional software in the form of betting bots. These will allow us to set the thresholds (a lower limit of 4.7 and an upper limit of 5.5). Even with the help of these betting bots we are faced with a dilemma. At what level should we set the stop loss. If we set it too high, say 5.1 then we risk always being in a position of having some liability if the horse wins, but the whole idea of this trading strategy is to have a win/win situation regardless of the outcome. Alternatively, if we set the stop loss too low, then we are taking a risk in the amount of our liability (who said betting exchanges were an easy way of making money?).

We also need to recognise how different betting bots treat stop losses. We have to distinguish between the trigger for an event and the event itself. Say we set a stop loss at 4.7, the trigger for the event would be if back odds =<4.7 then place bet to stakes indicated. However, there is one other variable that sits within this event, and that is the odds at which the bet is placed. There are two possibilities. The first of these is if back odds =<4.7 then place bet to stakes indicated @ 4.7. The second possibility is if back odds =<4.7 then place bet to stakes indicated @ 1.01.

What happens in the first case is that the betting bot will poll the betting exchange and when the back price reaches or is lower than 4.7 then it will place your bet requesting odds of 4.7. Should the odds have jumped from 4.8 to 4.6, then the bet will be placed, but if the price does not then go back out to 4.7 it will never get matched and you will be left in a total liability position.

In the second case the betting bot will poll the betting exchange and when the back price reaches or is lower than 4.7 it will place a bet at odds of 1.01. Using the concept of equitable betting, the betting exchange will match the bet at the highest available price, so your bet will always get matched (it may not be at 4.7, but you will never be left with a total liability).

To sum up: firstly we should always try and find a horse who we believe will not win. We will always place a lay bet first. We will set our spread at a low level two to three ticks. If we are using a betting bot (and this is really the only option) we will set a stop loss position, but the spread for this will be of a magnitude several times higher than our upward spread. We have to ensure that the betting bot understands the difference between the trigger for an event and the event itself.

Finally, we must always measure possible returns against liability to get a true understanding of the odds at which we are betting.